OmniVIL-Calib: Target-Free Joint Calibration for Omnidirectional Camera, IMU, and LiDAR

Publication: IEEE Robotics and Automation Letters (RA-L)

Project Overview / 项目概览

OmniVIL-Calib is an open-source toolbox designed for the joint extrinsic calibration of multiple sensors, specifically focusing on the integration of Omnidirectional Cameras, IMUs, and LiDARs. Unlike traditional methods, our approach is target-free, meaning it does not require specific calibration boards or controlled environments.

OmniVIL-Calib 是一个开源工具箱，专为多传感器的联合外参标定而设计，尤其专注于全景相机、惯性测量单元 (IMU) 和激光雷达 (LiDAR)的集成。与传统方法不同，我们的方法无需目标，这意味着它不需要特定的标定板或受控环境。

Key Contributions / 核心贡献

• Target-Free Calibration: Enables calibration in arbitrary natural environments by leveraging motion and feature distribution.
• Multi-Sensor Fusion: Provides a unified optimization framework for spatially and temporally aligning LiDAR, IMU, and omnidirectional cameras.

• Robustness: Demonstrated high accuracy and reliability across various complex indoor and outdoor robotic platforms.

• 无标定板标定： 利用运动和特征分布，可在任意自然环境中进行标定。
• 多传感器融合： 提供统一的优化框架，用于在空间和时间上对齐激光雷达、惯性测量单元和全景相机。
• 鲁棒性： 在各种复杂的室内外机器人平台上均展现出高精度和高可靠性。